Characteristics of Marrow Production and Reticulocyte Maturation in Normal Man in Response to Anemia

Characteristics of marrow production and reticulocyte maturation in normal man in response to anemia

Robert S. Hillman

J Clin Invest. 1969;48(3):443-453. https://doi.org/10.1172/JCI106001.

Research Article

Erythropoiesis in normal man was studied during periods of phlebotomy-induced anemia of varying severity. This study permitted a comparison of marrow production measurements over a wide range of marrow production levels. As long as the serum iron remained above 50 μg/100 ml, measurements of plasma iron turnover provided an excellent index of marrow production at all levels of red cell production. In contrast, the absolute reticulocyte count demonstrated a poor correlation with the other measurements. This was shown to be the result of a prolongation of the time required for circulating reticulocytes to lose their reticulum, which correlated with the severity of the anemia. For the clinical application of the reticulocyte count as a measurement of marrow production, an adjustment must be made for this alteration in the circulating reticulocyte maturation time.

Find the latest version: https://jci.me/106001/pdf Characteristics of Marrow Production and Reticulocyte Maturation in Normal Man in Response to Anemia

ROBERT S. HILLMAN From the Department of Medicine, University of Washington School of Medicine, Seattle, Washington 98104

A B S T R A C T Erythropoiesis in normal man was stud- normal human erythroid marrow to graded levels of ied during periods of phlebotomy-induced anemia of anemia maintained for 3-5 wk. This examination per- varying severity. This study permitted a comparison of mitted a comparison of erythropoietic measurements marrow production measurements over a wide range of over a wide range of marrow production levels and a marrow production levels. As long as the serum iron definition of the changes in reticulocyte maturation remained above 50 ug/100 ml, measurements of plasma which are important to the interpretation of the reticu- iron turnover provided an excellent index of marrow locyte count as a production measurement. production at all levels of red cell production. In contrast, the absolute reticulocyte count demonstrated a METHODS poor correlation with the other measurements. This was All studies were performed on the Clinical Research Center shown to be the result of a prolongation of the time re- of the King County Hospital, Seattle, Washington. Partici- quired for circulating reticulocytes to lose their reticu- pants included two patients with hemochromatosis, who re- lum, which correlated with the severity of the anemia. quired prolonged phlebotomy threapy, and six normal volun- For the clinical application of the reticulocyte count as teers. The subjects were males between the ages of 23 and 48. All studies were carried out according to the guidelines of a measurement of marrow production, an adjustment the Helsinki Declaration for participation of human volun- must be made for this alteration in the circulating retic- teers. Before study, each individual was carefully evaluated ulocyte maturation time. to exclude renal impairment or other significant illnesses which could have interfered with maximum erythropoietin INTRODUCTION output and marrow response to phlebotomy. During an ini- tial 10 day observation, base line studies of marrow func- Erythropoiesis in man has been studied by a number of tion were obtained, including two measurements of plasma techniques, including: the reticulocyte count, plasma iron iron turnover, a marrow iron transit time, serum iron, and turnover, 51Cr or diisopropylfluorophosphate (DF"P) total iron binding-capacity determinations, a 'Cr-labeled red red cell life-span, anid stercobilinogen excretion. Except cell mass, mean cell indices, repeated reticulocyte counts, and a bone marrow examination with Prussian blue stain for the reticulocyte count, these measurements are in for iron stores. Throughout the control and subsequent study essential agreement and apparently correlate with the periods, the subjects were kept on a high protein diet, and actual level of erythroid marrow production. However, folic acid was provided in excess, 5 mg twice a day orally. this conclusion is based primarily on measurements in In order to investigate the role of iron in the marrow production response, iron was provided by varying methods pathological states (1-5). Studies of the normal marrow when required (13). have been restricted to characterization of basal produc- After completion of base line studies, the subjects were tion and the acute response of marrow to sudden anoxia, phlebotomized, letting 500-1000 ml of whole blood to lower phlebotomy, or transfusion polycythemia (1-3, 6-10). the hematocrit to a level of 32-37%. The total blood volume As for the reticulocyte count, it has generally been dis- was kept constant by immediate infusion of an equal quantity of 5% albumin solution. Phlebotomies of more than 500 carded because of a poor correlation with the other pro- ml were accomplished in two stages by complete replace- duction measurements (11, 12). ment of the plasma volume after removal of each 500 ml. The present study has examined the response of the Over the next 3-5 wk, the hematocrit was maintained at this level by graded daily phlebotomy adjusted to remove Received for publication 20 June 1968 and in revised form enough red cells to compensate for increased production. 4 November 1968. An equal volume of 5%o albumin solution was infused with The Journal of Clinical Investigation Volume 48 1969 443 any phlebotomy of more than 50 ml of whole blood.' The to 0-4°C, centrifuged, and the separated plasma was counted erythroid marrow production response was then monitored in a gamma-well counter2 to 10,000 or more counts above with twice daily reticulocyte counts, measurements of the background for ± S1% accuracy. The plasma iron turnover plasma iron turnover at 7-day intervals, and from the level was calculated according to Bothwell and Finch (14) with of daily phlebotomy required to maintain the hematocrit at the mean of two serum irons drawn at 0 time and 30 min this level (see below, phlebotomy production measurement). after radioiron injection and the tj clearance. The radioiron With the hematocrit maintained at 32-37%, the erythroid administered to any single patient was limited to a maxi- marrow production response was measured in eight indi- mum of 25 uc for the entire study. viduals, six normals relying on either reticuloendothelial The plasma iron turnover production index was calculated cell iron stores and (or) orally administered iron, and two according to Giblett et al. (1) with 0.65 mg/100 ml of hemochromatosis patients with excessive parenchymal and whole blood per 24 hr as normal. reticuloendothelial cell iron stores. Patients on orally ad- Patient's PIT ministered iron received a 300 mg ferrous gluconate tablet PIT production index = -a0 65 at 2-hr intervals while awake for a total of 8-9 doses/day. Marrow production responses were monitored until produc- In order to study the effect of varying levels of reticulo- tion had plateaued for at least 10 days. This plateau was cytosis on the plasma iron turnover measurement, direct considered to represent a maximum level of marrow produc- reticulocyte uptake of radioiron was measured with each tion for the selected conditions of anemic stress and iron plasma iron turnover by in vitro incubation of the patient's supply. blood with '9Fe-labeled transferrin. Whole blood was col- Subsequently, six of the normal volunteers and one sub- lected in heparin, centrifuged, and the plasma removed. ject with hemochromatosis were studied at a hematocrit After adjusting the plasma iron to 200-250 ,tg/100 ml with level of 25-30%, maintained over a 3-5 wk period. This ferrous citrate-59Fe, sufficient plasma was recombined with level of phlebotomy-induced anemic stress was tolerated by the red cell fraction to obtain a hematocrit of 50-60%. the subjects without difficulty. One individual worked full Once a zero time sample was obtained the red cell suspen- time cleaning and steaming Alaskan King crab; the others sion was incubated in a water bath at 37°C. Samples were continued their normal daily activities without restriction. removed at 60 and 120 min, immediately chilled to 0-40C, centrifuged, and plasma aliquots obtained for counting. The Marrow production studies red cell fraction was washed three times in cold saline, Plasma iron turnover measurements. All determinations lysed in distilled water, and an aliquot counted for radio- of the plasma iron turnover were performed at 10:00 a.m. in activity. Red cell iron uptake was calculated from the se- a fasting state. With patients on orally administered iron rum iron and the net gain in activity of the red cell frac- the radioiron was injected 1 hour after the last oral dose. tion over time zero. 1-10 uc of ferrous citrate-"Fe was incubated with 5-10 ml Phlebotomy production measurements. Production was of the patient's plasma for 30 min before injection. The in- also estimated from the level of daily phlebotomy, because jected dose and standard dilution were determined gravi- if the hematocrit is kept at a constant level, red cell produc- metrically. Accurately timed venous samples were obtained tion must equal the amount of packed red cells removed at frequent intervals over the next 2 hr, immediately chilled by phlebotomy plus the amount of red cells dying each day. 1 5%o albumin solution, Albumisol (Merck, Sharp & Dohme, West Point, Pa.) was administered in equal volumes for any phlebotomy exceeding 50 ml of whole blood. Repeated blood volumes (51Cr-labeled red cell mass and (or) radioiron-labeled transferrin plasma volume) at weekly intervals confirmed the fact that this technique kept the blood volume constant throughout the periods of prolonged phlebotomy. Hct 32-377( Hct 25-30%

Week ..... I 2 3 4 5 6 7 8 9 Patient 1 5150 5200 5400 5210 5050 5320 5180 5060 2 3780 3800 3655 - 3850 4050 3780 3820 3650 3 4320 4550 4450 4300 4610 4400 - 4510 4480 'Nuclear-Chicago gamma-well counter, Nuclear-Chicago Corporation, Des Plaines, Ill. aThe amount of red cells lost from senescence each day (footnote 3 continued) was obtained by daily calculation of the red cell age distri- DED (day 1) = 1/120 X RCV. bution as continuously modified by the previous days' cell RCV (day 2) = RCV (day 1) - DED (day 1). death, actual cell removal by phlebotomy, and the change in DED (day 2) = F9 RCV (day 2)- REM production pattern. For example, on day 1 the red cell age RCV (day 2) was assumed to be random so that 1/120th of the red cell XTBV X Hct (day2) X 0.91 mass on that With removal of cells died day. by phlebotomy RCV (day 3) = RCV (day 2) - [DED (day 2) or test sampling, the age distribution was skewed by the appearance of a younger population at increased levels of RCV +REM (day 2) 1J' XXTBVxHct XO091 marrow production. Thus, on each subsequent day, the quantity of senescent cells was recalculated to adjust for DED (day 3) = 118 [RCV (day 3) - REM cell loss and to exclude the increased production of young RCV (day 3) ctc. cells by use of the following formulas: XTBV X Het (day 3) X 0.91 444 R. S. Hillman Therefore, for each period of constant anemia, total produc- iron turnover determination. Whole blood was obtained tion was estimated as the sum of the amount of red cells twice daily for 10 days after injection and 2 ml counted for removed plus the calculated amount of cells lost from radioactivity. We measured the transit time (days) as the senescence. Daily production was expressed in two ways: (a) time from 50% clearance of radioiron from the plasma to as production/day by relating the total production to the 50%o reappearance in red cells, assuming the 10-day level configuration of the reticulocyte curve (Fig. 1) and (b) as as 100%o reappearance. Measurements were performed in mean daily production for accumulated periods of not less the basal state and at each level of anemia once production than 7 days during periods of constant reticulocytosis. The had reached a plateau. production index equaled the mean daily production divided Reticulocyte maturation. Circulating reticulocyte matu- by 1/120th of the original red cell volume (the assumed ration was characterized by (a) classification of reticulum basal state production for the subject). content according to Heilmeyer (16) and (b) calculation For reasonable accuracy, this calculation required that of a circulating reticulocyte maturation time. (a) all new cells produced during the study have a life- (a) In two subjects, daily reticulocyte smears were graded span of at least 60 days and (b) the mean corpuscular hemo- for individual cell reticulum content. At least 100 reticulo- globin concentration of the new cells remains normal. The cytes were examined on each smear and graded on a scale latter was monitored with frequent red cell constants dur- of 0-IV, where type 0 = a typical orthochromatic normo- ing the studies. The first assumption was examined by blast or a cell which contains both nuclear fragments and comparing the final calculated life-span distribution to an large amounts of reticulum, I = a macrocytic cell nearly filled actual determination of the "Cr life-span of the circulating with a large amount of reticulum, II = a macrocytic cell red cells at the end of the phlebotomy study. 100 ml of whole containing a number of strands of reticulum, III = a normo- blood was obtained on the 50-70th day of the plebotomy cytic or slightly macrocytic cell with a few strands of reticu- study and labeled with 150 ,uc of "Cr in the presence of lum, and IV = a normocytic cell with a single strand or a acid citrate dextrose (ACD) solution. After washing the few fragments of reticulum. cells with sterile saline, one-third of the original sample (b) In view of the progressive reduction in transit time was injected into each of three normal recipients. Blood and increase in amount of reticulum in the reticulocytes of samples were obtained three times a week for the next 7 wk anemic individuals, the discrepancy between the absolute to determine the "Cr disappearance rate. Red cell life- reticulocyte counts and other measurements of production was spans were expressed both as the tj disappearance of "Cr assumed to represent in large part a prolongation of the and as the percentage of labeled red cells still circulating time required for reticulum to disappear from the circulating at 50 days after correction for a 1% "Cr elution rate. reticulocyte, the circulating reticulocyte maturation time. Reticulocyte production. Reticulocyte counts were per- This assumption is supported by experiments in animal formed twice daily throughout the studies. After incubation models in which in vivo studies of maturation of reticulum, with new Methylene blue, cover slip smears were counter- ribosomes, and the ability to synthesize hemoglobin have stained, mounted in pairs, and at least 5000 consecutive red been carried out (reference 18 and footnote 4). A circulating cells were enumerated for each time point by two obser- reticulocyte maturation time was therefore calculated at vers. The observed reticulocyte counts were corrected for each hematocrit level from the absolute reticulocyte counts variations in hematocrit to obtain an absolute reticulocyte and the plasma iron turnover and (or) phlebotomy produc- percentage: tion measurements. Absolute reticulocyte count (%) Circulating reticulocyte maturation time (days =Observed reticulocyte count (%) X Observed Hct Absolute reticulocyte count (%) Plasma iron turnover or phlebotomy production index. Marrow and reticulocyte maturation studies Normal maturation time (days) = 1.0. Marrow iron transit times. The marrow iron transit Throughout all studies frequent measurements of red cell time was measured after injection of 'Fe for the plasma constants, serum iron, and total iron binding capacity (18, (footnote 3 concluded) 19), smear morphology, and serum proteins were obtained. DED = Cell dying in next 24 hr. RCV = Red cell volume (ml) as measured RESULTS by 5"Cr-labeled red cell mass. were RCV (day 2, 3, etc.) = That portion of the red cell mass Eight subjects studied both in the normal state and remaining after phlebotomy and after induction of specific levels of anemia by graded the previous days' cell death. phlebotomy. The marrow production response was deter- REM = Red cells (ml) removed by phle- mined from serial measurements of the plasma iron turn- botomy. over, reticulocyte counts, and the level of daily phle- TBV = Total blood volume (ml) as deter- botomy, as shown for one of the subjects (Fig. 1). This mined from the 5"Cr-labeled red cell mass and plasma volume determination permitted a comparison of the three mea- measurements of the transferrin- surements of production as well as characterization of "Fe clearances. reticulocyte behavior with increasing anemia. Hct = Hematocrit (%) microhematocrit method. 'Ganzoni, A., R. S. Hillman, and D. Heywood. Hemo- 0.91 = Total body hematocrit: venous he- globin synthesis during in vivo maturation of the rat reticu- matocrit ratio ( 15). locyte. Data to be published Marrow Production and Reticulocyte Maturation in Normal Man 44,5 E

-= I- 6 o4 48 2 , 2 .~~~395mI~ ~ ~ ~ ~ 50 l__'

+ 4 + + + + + a: 1.0 1.0 2.4 2.6 2.75 3.5 3.2 3.7 0 10 20 30 40 50 60 70 80 Days FIGURE 1 Serial measurements of marrow production in one subject during a control period and after the hematocrit was lowered to 32-37%o and subsequently 25-30%o for 3-5 wk. The hematocrit was kept constant by daily graded phlebotomy. (REM = amount of blood removed each day as packed red cells.) Cell senescence (DED) was calculated daily (see Methods) and added to the amount of cells removed to obtain a phlebotomy production measurement. This was expressed as mean daily production for periods of constant reticulocytosis (bar graphs indicating both milliliters of cells produced each day and production index) and by smoothing the total production for each hematocrit level to the shape of the reticulocyte curve (solid line under the reticulocyte curve). Measurements of plasma iron turnover were performed at 7- to 10-days intervals and are expressed as the PIT index. Reticulocyte counts (closed circles) were obtained twice daily and corrected for hematocrit variation. Measurements of marrow production these anemic conditions have both a life-span greater Phlebotomy production measurement. As seen in than 60 days and a normal mean corpuscular hemoglobin Fig. 1, the amount of cells removed by phlebotomy concentration (MCHC). The first requirement was ex- a determination on (REM) was continuously adjusted as marrow production amined by performing 'Cr life-span increased to maintain the hematocrit level first between the cells at the end of the study (Fig. 2). After 65 days 32 and 37% and subsequently between 25 and 30%. of continuous phlebotomy, the calculated red cell age Daily levels of marrow production were then estimated distribution for the subject shown in Fig. 1 was mark- from the amount of packed red cells removed by phle- edly shifted to a younger aged population; 93% of the botomy plus the estimated amount of cells lost from circulating cells were less than 60 days old. When this senescence (see Methods). To perform this calculation population was labeled with 'Cr and their life-span de- it was essential that all new red cells produced under termined in three normal recipients, Cr ti disappear- 446 R. S. HiUman too 90 80 70I 100 60i 50 A 40 *~~~~~ 30

100 I. I 88I *- 0- 70 70~~~~~~~ 60 0~~~~ > B 50 .- C-) 40 0 0) ._ 30 i 7%/O L- I 20 I 100 90 1NA 80 0. 70 C 60 50 40D 3C --- I .:00~~~~~0- L 20 40 60 80 100 0 10 20 30 40 50 60 DAYS DAYS FIGURE 2 Red cell age distribution and red cell life-spans after 65 days of continuous phlebotomy for the subject depicted in Fig. 1. The red cell age distribution was calculated from the phlebotomy data and the combined measurements of marrow production with the assistance of a digital computer. If a uniform age distribution on day 1 is assumed, the change in age distribution was calculated on each successive day of phlebotomy taking into account cell loss from senescence and phlebotomy and increased cell entry by way of increased red cell production. By the 65th day, it was calculated that 93% of the circulating red cell population was less than 60 days of age. When these cells were labeled with 51Cr, tj disappearance of 50, 55, and 57 days were observed after administration to three normal recipients. After correction for a 1% 'Cr elution rate (broken lines), 85-90% of the original labeled cells were still circulating in these recipients after 50 days. This confirmed the fact that the predicted shift to a younger population had occurred and that these cells had a relatively normal life-span. ances of 50, 55, and 57 days were observed. If a 1% 'Cr normal, the plasma iron turnover indices were in close elution rate is assumed, this indicated that 85-90% of agreement with the phlebotomy production measure- the labeled cell population was still circulating at 50 days, ments (Fig. 4). On only four occasions was there a a fact confirming the viability of the anemia-induced major discrepancy between the two measurements. In erythrocytes. Similar results for three other subjects each instance, the plasma iron turnover was performed are shown in Table I. As for the second requirement, at a time when the serum iron values were less than serial measurements of mean-cell constants were per- 50 ug/I100 ml. formed during the studies (Fig. 3). Although the mean Reticulocyte iron uptake was studied in conjunction corpuscular volume (MCV) and mean corpuscular he- with the measurements of plasma iron turnover (Table moglobin content (MCH) rose slightly at hematocrit II). At higher levels of marrow production, when the 25-30%, the MCHC did remain constant. Both findings reticulocyte counts were 13.0-15.0%, in vitro red cell up- supported the validity of the phlebotomy production mea- take increased to a maximum of 0.52 ug/ml red cells af- surement. This was further supported by the excellent ter a 120 min incubation. If it is assumed that in vivo correlation with the plasma iron turnover measure- reticulocyte uptake -would be of a similar magnitude, ments (see below). it may be estimated that at the highest levels of mar- Plasma iron turnover measurements. Over a wide row production with plasma iron turnovers of 8-10,000 range of marrow production levels, one to five times pg every 2 hr less than 500 pg was directly incorporated Marrow Production and Reticulocyte Maturation in Normal Man 447 TABLE I 25% grade III reticulocytes, the percentage of grade III 51Cr Life-Spans of Cell Populations after 50-65 days reticulocytes increased to 40% at hematocrit of 35 and of Continuous Phlebotomy 60-70% at hematocrit of 25-27%. At the lower hematocrit level, approximately 5-10% of grade II reticulocytes Calculated were also seen. age dis- Labeled tribution Both findings imply a premature delivery into cir- cells re- of cells culation of marrow reticulocytes which require a longer maining less than than normal period of maturation in circulation to lose 51Cr tq at 50 60 days Patient disappearance days* age their reticulum. Since this could in large part explain the discrepancy between the reticulocyte counts and production indices, approximate circulating reticulocyte A 50 85 93 maturation times were calculated from the absolute 55 88 57 90 reticulocyte counts and plasma iron turnover production indices at varying levels of marrow production and he- B 53 86 91 matocrit depression. As shown in Fig. 8, when these 60 92 reticulocyte maturation times were compared to the 15 hematocrit level, they progressively lengthened with C 45 79 86 increasing anemia. From a normal value of 0.8-1.2 days, 55 84 D 52 83 89 56 85 Control Hct 35-37 Hct 25-27 20- 0 0 * Assuming 1 % elution of 51Cr. 16- ? A S 12- into circulating reticulocytes. The primary distribution *000 aR 0 8 8- 00. It of injected radioiron continued to be nucleated marrow S 0 O 4- precursors. 0 Reticulocyte production measurements. With ane- I & *0 mia, the absolute reticulocyte counts (the observed 10' reticulocyte counts corrected for hematocrit variations) 0 11- were consistently higher than actual marrow production A (Figs. 1 and 5). This discrepancy was not a function 0 of the level of marrow production; the absolute reticulo- ; 90- .0 cyte counts exceeded the plasma iron turnover or phle- I 0 botomy balance indices by a factor of 1.5-3.0 both when v production was two to three times normal and at levels 341 I ~~~~oj 0 06 of five times normal 5). do (Fig. *00 4p 0@ 0 0 Marrow and reticulocyte maturation studies jS30~ I. Ir were 3281 Measurements of the marrow iron transit time CIAO V* made at each hematocrit level (Fig. 6). A progressive shortening of the iron transit time occurred as the se- I Du ol mei 01 0 verity of the anemia increased. By the time the hemato- E 32. ,° ° ° 1 .1 0 crit was reduced to 25-30% the marrow iron transit X 8 time had shortened to 1.5 days as compared to a normal time of 3.5 days. This was associated with the appearance in circulation of large, polychromatic reticulocytes containing greater than normal amounts of reticulum, mar- 0 2 4 6 8 10 row reticulocytes. When new Methylene blue-stained WEEKS reticulocyte smears were graded according to the Heil- FIGURE 3 Mean cell constants of four subjects during phle- meyer classification (16) it was apparent that the num- botomy to hematocrits of 32-37% and 25-30%o. The mean corpuscular volume (MCV) and mean corpuscular hemo- ber of cells containing increased amounts of reticulum globin content (MCH) increased slightly, while the mean correlated with the hematocrit level (Fig. 7). Whereas corpuscular hemoglobin concentration (MCHC) remained normal subjects demonstrated 75% grade IV and only constant throughout the studies. 448 R. S. Hiliman A4C

E 4.0 lq3 12.C X * /

o0 D S Il.ac . 0)~~~~~~~~~~~ 30 0)~~~~~~~~~~ IQC 0 3.0X 0 bc

1 .0 0 0 &

37.C e 0

0 a00 0 0 1.0 2.0 3.0 4.0 5.0 0 0 0 00 Phlebotomy production measurement (x's normal) 0 000 FIGURE 4 A comparison of the plasma iron turnover in- a 0 dices with the phlebotomy production measurement. When 0 measurements of the plasma iron turnover were performed at a time when the serum iron was less than 50 /Ag/100 ml (open circles), the plasma iron turnover gave a falsely low measure of production. *0 2': the circulating reticulocyte maturation time lengthened LO to 1.5-2 days at hematocrit of 32-37% and 1.7-3 days at hematocrit of 25-30%. 1.0 TO &O 4.0 5.0 &O 7.0 &U The prolongations of the maturation time of the cir- Marrow Production (x's normal) culating reticulocyte were also compared to the mea- FIGURE 5 A comparison of the absolute reticulocyte counts surements of marrow iron transit time. If the reticulocyte and plasma iron turnover (closed circles) or phlebotomy maturation prolongation were merely a reflection of pre- production indices (open circles) demonstrated a poor cor- mature delivery to circulation of otherwise normal relation. The absolute reticulocyte count exceeded marrow production by one and one-half to three times normal over a reticulocytes, a comparable shortening of the marrow wide range of production levels. iron transit time might be expected for increasing prolongations of circulating reticulocytes maturation. As shown in Fig. 9, when the transit time shortened from days. In only two subjects, the patients with hemo- 3.5 to 2.5 days there was a comparable prolongation chromatosis, did a comparable prolongation of the cir- of reticulocyte maturation. However, as the transit time culating reticulocyte maturation time occur for the lower fell to below 2 days the circulating reticulocyte matura- values of marrow iron transit time. tion time in most subjects remained between 1.5 and 2 DISCUSSION TABLE I I Whereas past studies of the human erythroid marrow's Reticulocyte Uptake of Radioiron response to anemia have been largely limited to pathological states (1-5), the present measurements of mar- Number row production were carried out in normal individuals of de- Iron uptake during prolonged periods of phlebotomy-induced anemia. Reticulocyte termina- count tions 60 min 120 min The results would support the previous conclusion that the plasma iron turnover reflects the level of erythroid % ug/ml red cells marrow activity (6-10, 13). Over a wide range of mar- 0.5- 1.5 3 0 0-0.12 row production levels, there was excellent correlation 3.5- 5.0 5 0-0.05 0-0.21 5.0-10.0 4 0.15-0.30 0.15-0.50 between the plasma iron turnovers and the marrow 10.0-15.0 6 0.05-0.42 0.18-0.52 production level as estimated from the level of daily phlebotomy required to maintain the hematocrit at a

Marrow Production and Reticulocyte Maturation in Normal Man 449 1.0 So

0 15 g0 0;d Co 0 E2f 0 0 E ._ 25 0 * S1.0%252.0 ./ a 0 h. 00 ._ . A 0o I maC a o o- S.-- 1.5 0 45 40 35 30U 25 0 ox Hemotocrit (%) cr1.0I FIGURE 6 Marrow iron transit times were measured at hematocrit levels ranging from 45 to 25%. As the hematocrit was lowered the transit time shortened from 3.5 to 1.7 days. 45 40 35 30 25 constant level. Only when the serum iron fell below Hemotocrit (%/6) 50 pg/100 ml did the measurements of plasma iron turn- FIGURE 8 Progressive prolongation of the circulating re- over fail to accurately reflect marrow production. In ticulocyte maturation time occurred with depression of the these situations, production apparently continued at hematocrit. If we assume that the discrepancy between the absolute reticulocyte count and plasma iron turnover is primarily related to a premature delivery of marrow reticulocytes which then require a longer time to lose their reticulum, the reticulocyte maturation time was calculated as: Reticulocyte maturation time Absolute reticulocyte count (days) Production index Closed circles, calculated from the plasma iron turnover; Open circles, calculated from phlebotomy production measurements.

levels of two to three times normal at a time when measurements of plasma iron turnover were less than twice normal. In part, this may reflect the fact that the plasma iron turnover only measures iron delivery to the marrow at one point in time and does not necessarily indicate total iron supply over a 24 hr period. However, the discrepancy may also imply a mechanism of direct trans- fer of iron from reticuloendothelial cells to marrow precursors, which becomes important at lower serum iron values. When the absolute reticulocyte counts were compared to the measurement of marrow production by plasma 0 10 20 30 40 50 60 70 iron turnover, the reticulocyte counts were consistently Days higher than actual marrow production over a wide range FIGURE 7 Reticulocytes were graded for their relative con- of production levels (Fig. 5). At the same time, large tent of reticulum according to the Heilmeyer classification numbers of polychromatic macroreticulocytes appeared (grade IV, closed circles; grade III, open circles; grade II, circled dots). As the hematocrit was lowered large num- in circulation. Both the number of these cells and their bers of reticulocytes containing increased amounts of reticu- relative reticulum content progressively increased as the lum, grade III and II types, appeared. hematocrit was lowered. Reiff, Nutter, Donohue, and 450 R. S. Hillman reticulocyte maturation. In small animals, Ambs (22) and Brecher and Stohlman (23) have demonstrated the occurrence of skipped mitoses at the basophilic or polychromatic normoblast stage which results in the re- 0 population of the circulation with macrocytes containing I- i greater than normal amounts of hemoglobin. Since such .2C cells may be short-lived (24-26), this phenomenon is of potential importance in the interpretation of the

o 0 reticulocyte count, both in terms of prediction of the maturation time of the circulating reticulocyte and the 0o correlation of the reticulocyte count with total marrow production. However, Ganzoni, Hillman, and Finch (17) have recently shown that these cells do circulate for at least one-third of their normal life-span so as not to in- terfere with a full expression of the phenomenon of reticulocyte maturation time prolongation. Moreover, in 3.0 25 2.0 1.5 normal man, there is little evidence for a significant Transit time (days) reduction in the number of mitoses within the marrow. The increases in the MCV and MCH in the FIGURE 9 Prolongation of the circulating reticulocyte matu- present ration time (closed circles, calculated from plasma iron turn- studies were not of a magnitude which would implicate overs; open circles, calculated from phlebotomy production skipped mitosis as an important mechanism in man dur- measurement) is compared to the shortening of the mar- ing periods of moderate anemia. Furthermore, when row iron transit time. At a hematocrit of 25 the transit time life-spans of the phlebotomy-induced red cell populations reduction was not accompanied by an equal prolongation of the reticulocyte maturation time in the normal subjects. Only were determined, a significant number of short-lived the patients with hemochromatosis demonstrated a com- cells was not observed. Of course, this measurement parable prolongation. would not have detected a loss of cells with a life-span of only a few days. However, this type of wastage seems Finch (20) and Gordon, LoBue, Dornfest, and Cooper unlikely. Not only have studies of small animals failed (21 ) have shown in the rabbit and rat that marrow to demonstrate an appreciable loss of macroreticulocytes reticulocytes are released from the marrow prematurely within the first few days (17), but also any major loss in response to anemia or erythropoietin stimulation. of short-lived cells in the present studies should have Ganzoni, Hillman, and Finch (17) have further demon- been reflected by a major discrepancy between the strated that these rat macroreticulocytes have a prolonged plama iron turnover and the phlebotomy production maturation time in vivo. This is accompanied by re- measurement. For these reasons, the present studies duction in the marrow iron transit time, an approximate would indicate that in man, at least, the high levels of measurement of the time required for basophilic nor- marrow production and changes in marrow maturation moblasts to mature to marrow reticulocytes and be which are induced by moderate anemia do not result delivered into circulation. Similarly, in man, the appear- in an abnormal reticulocyte product in which changes ance of increasing numbers of macroreticulocytes con- in reticulocyte maturation or life-span will invalidate taining greater than normal amounts of reticulum is as- the use of the reticulocyte count as a measurement of sociated with a progressive shortening of the iron transit effective marrow production. time. This would suggest that in large part the dis- The prolongation of the circulating reticulocyte matu- crepancy between the absolute reticulocyte count and ration time reflects the severity of the anemia. When actual marrow production levels is the result of a pro- circulating reticulocyte maturation times were calcu- longation of the circulating reticulocyte maturation lated from the absolute reticulocyte counts and plasma time. Since prematurely delivered marrow reticulocytes iron turnover measurements, progressive prolongation require an extra 1 or 2 days to lose their increased of the maturation time occurred as the severity of the amounts of reticulum, they tend to accumulate in circu- anemia increased (Fig. 8). This was accompanied by a lation and falsely elevate the reticulocyte count. shortening of the marrow iron transit time. At a hemato- Before the concept of a prolongation of the circulating crit of 32-37%, the transit time shortened commensurate reticulocyte maturation time is accepted as the lone with the prolongation of the reticulocyte maturation so cause of the discrepancy between the absolute reticulo- that the total maturation time, the sum of the two mea- cyte counts and other measurements of marrow produc- surements, remained constant. At the 25-30% hemato- tion, it is important to exclude other disturbances in crit level, further shortening of the marrow iron transit Marrow Production and Reticulocyte Maturation in Normal Man 451 time was not necessarily reflected in equal prolongation ACKNOWLEDGMENTS of the reticulocyte maturation. In this regard, it should I gratefully acknowledge the assistance of Doctors Peter be noted that there was a concomitant prolongation of Ways and Clement Finch for their interest and advice and reticulocyte maturation for the degree of iron transit Dr. Paul Worrell, Mrs. Maura Oakes, Mrs. Mudite Peter- time reduction in both patients with hemochromatosis. sons, and Mrs. Caryl Finholt for their technical help. However, in those individuals relying on orally admin- This work was supported in part by U. S. Public Health Service Grant AM 09950. A portion of the work was con- istered iron or reticuloendothelial iron stores, the total ducted through the University of Washington Clinical Re- maturation period fell from a normal value of 4.3-4.5 to search Center at King County Hospital, supported by the 3.4-4.0 days. In view of the limitation of iron supply National Institute of Health (Grant FR-133). in the latter subjects, relative iron deficiency may have caused either a reduction in the total reticulocyte matu- REFERENCES ration period or a change in the distribution of radio- 1. Giblett, E. R., D. H. Coleman, G. Pirzio-Biroli, D. M. iron to marrow precursors so as to inordinately shorten Donohue, A. G. Motulsky, and C. A. Finch. 1956. Eryth- the transit time. This theory is supported by rokinetics. Quantitative measurements of red cell pro- previous duction and destruction in normal subjects with anemia. studies of pernicious anemia patients during vitamin Bn Blood. 11: 291. therapy (27). in wXNhich the iron supply was well main- 2. Huff, R. L., T. G. Hennessy, R. E. Austin, J. F. Garcia, tained during studies of their marrow maturation char- B. M. Roberts, and J. H. Lawrence. 1950. Plasma and acteristics. In this situation, measurements of the mar- red cell iron turnover in normal subjects and in patients having various hematopoietic disorders. J. Clin. Invest. row iron transit time and reticulocyte maturation time 29: 1041. at hematocrit level of 15% demonstrated a comparable 3. Huff, R. L., P. J. Elmlinger, J. F. Garcia, J. M. Oda, shortening of the transit time for a prolongation of M. C. Cockrell, and J. H. Lawrence. 1951. Ferrokinetics reticulocyte maturation of 2.3-2.5 days. in normal persons and in patients having various erythropoietic disorders. J. Clin. Invest. 30: 1512. The concept of a progressive prolongation of the cir- 4. Finch, C. A., D. H. Coleman, A. G. Motulsky, D. M. culating reticulocyte maturation time with increasing Donohue, and R. H. Reiff. 1956. Erythrokinetics in levels of erythropoietin stimulation is an important con- pernicious anemia. Blood. 11: 807. sideration in interpreting the reticulocyte response in 5. Sturgeon, P., and C. A. Finch. 1956. Erythrokinetics in pathological states. Only by correcting the absolute Cooley's anemia. Blood 12: 64. 6. Bothwell, T. H., A. V. Hurtado, D. M. Donohue, and reticulocyte count for the appropriate prolongation of C. A. Finch. 1957. Erythrokinetics. IV. The plasma iron circulating reticulocyte maturation can a reasonable turnover as a measure of erythropoiesis. Blood. 12: 409. index of production be obtained. Although it is pres- 7. Lawrence, J. H., R. L. Huff, W. Siri, L. R. Wasserman, ently impossible to express this prolongation as a func- and T. G. Hennessy. 1952. A physiological study in the tion of measured erythropoietin levels, it is possible to Peruvian Andes. Aqta Med. Scand. 142: 117. 8. Reynarfarje, C., R. Lozano, and J. Valdiveisco. 1959. estimate a maturation time for a specific level of hemato- The polycythemia of high altitudes: iron metabolism and crit depression in those states in which erythropoietin related aspects. Blood. 14: 433. stimulation appears appropriate for the severity of the 9. Lawrence, J. H., P. J. Elmlinger, and G. Fulton. 1952. anemia (28). Thus, as the hematocrit is reduced to Oxygen and the control of red cell production in pri- approximately 35%', the maturation time of the circu- mary and secondary polycythemia. Effects on the iron turnover patterns with Fe' as tracer. Cardiologia. 21: lating reticulocyte increases from 1.0 to 1.5-1.7 days. 337. With more severe anemia the maturation time will in- 10. Pollycover, M., and R. Mortimer. 1961. The quantitative crease to 1.7-2.0 days at a hematocrit of 20-30% and to determination of iron kinetics and hemoglobin synthesis as much as 2.3-2.5 days at a hematocrit below 20% in human subjects. J. Clin. Invest. 40: 753. (27). This estimated maturation time may then be em- 11. Cline, M. J., and N. I. Berlin. 1963. The reticulocyte count as an indicator of the rate of erythropoiesis. Amer. ployed to calculate a corrected reticulocyte count and J. Clin. Pathol. 2: 121. production index by using the formula: 12. Crosby, W. H., C. A. Finch, F. I. Haurain, F. Stohlman, and L. Wasserman. 1961. Modern methods for quanti- Corrected, absolute reticulocyte count (%) tation of red blood cell production and destruction: Absolute reticulocyte count erythrokinetics. Blood. 18: 225. 13. Hillman, R. S., and P. Henderson. 1969. The control of maturation time (days) marrow production by the level of iron supply. J. Clin. index Invest. 48: 000. Reticulocyte production 14. Bothwell, T. H., and C. A. Finch. 1962. Iron Metabo- Corrected reticulocyte count 5 lism. Little, Brown and Company, Boston. 1.0 15. Mollison, P. L. 1967. Blood Transfusion in Clinical Medicine. F. A. Davis Co., Philadelphia. 'The normal reticulocyte count and reticulocyte index is 16. Heilmeyer, L. 1931. Blutfarbstoff wech sel studien. 1.0 (27). Deut. Arch. Klin. Med. 171: 121. 452 R. S. Hillman 17. Ganzoni, A., R. S. Hlillman, and C. A. Finch. 1969. 23. Brecher, G., and F. Stohlman, Jr. 1961. Reticulocyte Maturation of the macroreticulocyte. Brit. J. Haematol. size and erythropoietic stimulation. Proc. Soc. Exp. Biol. 16: 119. Med. 107: 887. 18. Bothwell, T. H., and B. Mallett. 1955. The determination 24. Stohlman, F., Jr. 1961. Humoral regulation of erythro- of iron in plasma or serum. Biochem. J. 59: 599. poiesis. VII. Shortened survival of erythrocytes produced 19. Morgan, E. H., and G. Carter. 1960. Plasma iron and by erythropoietin or severe anemia. Proc. Soc. Exp. Biol. iron binding capacity levels in health and disease: with Med. 107: 884. an improved method for estimation of plasma iron con- 25. Berlin, N. J., and C. Lotz. 1951. Lifespan of the red centration and total iron binding capacity. Australas. blood cell of the rat following acute hemorrhage. Proc. Ann. Med. 9: 209. Soc. Med. 78: 788. 20. Reiff, R. H., J. Y. Nutter, D. M. Donohue, and C. A. Exp. Biol. Finch. 1958. The relative number of marrow reticulo- 26. Card, R. T., and L. S. Valberg. 1967. Characteristics cytes. Amer. J. Clin. Pathol. 30: 199. of shortened survival of stress erythrocytes in the rabbit. 21. Gordon, A. S., J. LoBue, B. S. Dornfest, and G. W. Amer. J. Physiol. 213: 566. Cooper. 1962. Reticulocyte and leukocyte release from 27. Hillman, R. S., J. Adamson, and E. R. Burka. 1968. isolated perfused rat legs and femurs. In Erythropoiesis. Characteristics of the vitamin B12 correction of the ab- Grune & Stratton, Inc., New York. 321. normal erythropoiesis of pernicious anemia. Blood. 21: 22. Ambs, E. 1960. Uber Grosse Und ham oglobinreiche 419. reticulocytin inkerkrise von aneinien. Schweiz. Med. 28. Hillman, R. S., and C. A. Finch. 1967. Erythropoiesis, Wochenschr. 90: 413. normal and abnormal. Seminars Hematol. 4: 327.

Marrow Production and Reticulocyte Maturation in Normal Man 453